Problem with Bowtie alignment

[Linas]

It appears that bowtie do not print out U nucleotides. However, could it be the problem of low mapping percentage that input is sequence data (that is cDNA) and reference is mature miRNA?

[Linas]

Novice here. I tried to align miRNA sequencing data (Illumina) with mature miRNA from mirbase.org database using Bowtie. Reads are adapter trimmed and mapped to human genome hg19 by someone else. Length is about 26 nt. However, i get very small portion of aligned reads to mature miRNAs (less than 1%). I have tried options: -n 0 -l 20 (-l 17 - no difference). However, even those reads that got aligned were weird because of the length of the aligned reads in an output.

I then experimented with some mirna sequences to see how do results look in a more controlled manner: i aligned sequence from reference genome (mature miRNAs) with a indexed mature miRNAs reference. Of course, it got aligned but in a output it wrote some 13 nucleotide sequence that was not in input.
Input was 22 nt:
>hsa-let-7a-5p MIMAT0000062 Homo sapiens let-7a-5p
UGAGGUAGUAGGUUGUAUAGUU.

Output was 13 nt:
hsa-let-7a-5p MIMAT0000062 Homo sapiens let-7a-5p + hsa-let-7a-5p 0 GAGGAGAGGGAAG IIIIIIIIIIIII 0

I would like to know what this output means. As far as i know the sequence should be the sequence that got aligned. If i trimm (-3 <int> or -5 <int) the 13 nt sequence shortens by a indicated number.